On Robust Mean Estimation under Coordinate-level Corruption
This addresses robust statistical estimation for data with coordinate-level corruptions, which is incremental as it builds on prior adversary models with a new measure and practical algorithms.
The paper tackles robust mean estimation under coordinate-level corruption by introducing a Hamming distance-based measure for more realistic adversary models, and shows that a two-step data cleaning approach outperforms structure-agnostic methods, achieving accurate estimation even with high-magnitude corruption.
We study the problem of robust mean estimation and introduce a novel Hamming distance-based measure of distribution shift for coordinate-level corruptions. We show that this measure yields adversary models that capture more realistic corruptions than those used in prior works, and present an information-theoretic analysis of robust mean estimation in these settings. We show that for structured distributions, methods that leverage the structure yield information theoretically more accurate mean estimation. We also focus on practical algorithms for robust mean estimation and study when data cleaning-inspired approaches that first fix corruptions in the input data and then perform robust mean estimation can match the information theoretic bounds of our analysis. We finally demonstrate experimentally that this two-step approach outperforms structure-agnostic robust estimation and provides accurate mean estimation even for high-magnitude corruption.